MiMo-V2-Pro
XiaomiProprietary
Description
MiMo-V2-Pro is Xiaomi's flagship foundation model built for real-world agentic workloads. A Mixture-of-Experts model with over 1T total parameters and 42B active, roughly 3x larger than MiMo-V2-Flash. Inherits the Hybrid Attention mechanism with a 7:1 ratio and supports up to 1M-token context. Optimized for coding and agent tasks through post-training scaling across diverse agent scenarios, with native support for complex tool calling and multi-step reasoning.
Release Date
2026-03-18
Parameters
1.0T
Context Length
1.0M
Modalities
text
Capability Radar
45
general
42
coding
87
reasoning
60
scienceest.
80
agents
0
multimodal
Science uses a reasoning proxy when dedicated science benchmarks are unavailable.
Rankings
| Domain | #Rank | Score | Source |
|---|---|---|---|
| Agents & Tools | 32 | 63.0 | LS |
| Code Ranking | 53 | 72.0 | AA |
| General Ranking | 30 | 84.0 | AA |
| Science | 38 | 75.0 | AA |
Benchmark Scores (LLM Stats)
Agents
GDPval-AA
1426.00 / 3000SR
DeepSearchQA
86.7%SR
PinchBench
81.0%SR
Claw-Eval
61.5%SR
Terminal-Bench 2.0
57.1%SR
Code
SWE-Bench Verified
78.0%SR
SWE-bench Multilingual
71.7%SR
Communication
Tau2 Telecom
96.8%SR
AA Evaluation Indices
Intelligence Index49.2
Coding Index41.4
Tau20.9
Gpqa0.9
Ifbench0.7
Lcr0.6
Scicode0.4
Terminalbench Hard0.4
Hle0.3
LLM Stats Category Scores
Finance100
General100
Legal100
Agents100
Reasoning100
Communication100
Search90
Tool Calling80
Frontend Development80
Code70
Coding70
Pricing
Input Price$1 / 1M tokens
Output Price$3 / 1M tokens
Blended Price (3:1)$1.5 / 1M tokens
Speed
Tokens/sec65.2 tokens/s
Time to First Token2.11s
Time to Answer46.98s
Available Providers
(LS internal units)| Provider | Input Price | Output Price |
|---|---|---|
| Xiaomi | 1.0M | 3.0M |